Michael Lu

About

I am a final year PhD candidate at Simon Fraser University under the supervision of Dr. Mo Chen and Dr. Sharan Vaswani. I’m broadly interested in Safe Reinforcement Learning. In particular, I’m interested in developing principled algorithms for Constrained Markov decision processes with last-iterate guarantees. I'm also interested in exploring how these methods can then be applied to Reinforcement Learning with Preferences methods.

Prior to this, I completed a B.Sc. (Joint Honours) in Mathematics and Computing Science at Simon Fraser University. In my spare time, I enjoy cycling around Vancouver.

Publications

2026

• Augmented Lagrangian Method for Last-Iterate Convergence for Constrained MDPs

  Michael Lu, Max Qiushi Lin, Mo Chen, and Sharan Vaswani

  NeurIPS 2026

  Paper

2025

• Safe Learning in the Real World via Adaptive Shielding with Hamilton-Jacobi Reachability

  Michael Lu, Jashanraj Gosain, Luna Sang, and Mo Chen

  L4DC 2024

  Paper

• Rethinking the Global Convergence of Softmax Policy Gradient with Linear Function Approximation: The Case of Multi-Armed Bandits

  Max Qiushi Lin, Jincheng Mei, Matin Aghaei, Michael Lu, Bo Dai, Alekh Agarwal, Dale Schuurmans, Csaba Szepesvári, and Sharan Vaswani

  Preprint

  Paper

• Fast Confidence-Aware Human Prediction via Hardware-accelerated Bayesian Inference for Safe Robot Navigation

  Michael Lu, Minh Bui, Xubo Lyu, and Mo Chen

  Preprint

  Paper

2024

• Towards Principled, Practical Policy Gradient for Bandits and Tabular MDPs

  Michael Lu, Matin Aghaei, Anant Raj, and Sharan Vaswani

  RLJ 2024

  Paper Code

2021

• Real-time Hamilton-Jacobi Reachability Analysis of Autonomous System With An FPGA

  Minh Bui, Michael Lu, Reza Hojabr, Mo Chen, and Arrvindh Shriraman

  IROS 2021

  Paper